Skip to main content
Fig. 4 | BMC Bioinformatics

Fig. 4

From: SweepCluster: A SNP clustering tool for detecting gene-specific sweeps in prokaryotes

Fig. 4

Comparison of the LD patterns for the SNPs before and after pre-selection for the genotype datasets of S. pyogenes (AC) and S. suis (DF). A, D The LD pattern of SNPs in the most significant cluster for all segregating SNPs from S. pyogenes and S. suis, respectively. B, E The LD pattern of the selected SNPs with phenotypic association in S. pyogenes and population differentiation in S. suis. C, F Distribution of the average level of inter-SNP LD in the clusters for all segregating SNPs and the selected subset of SNPs from S. pyogenes and S. suis, respectively. The LD pattern in (A) involves 1,014 SNPs located in the genomic region 1,273,267–1,286,739 of S. pyogenes AP53. The pattern in (B) involves the same set of SNPs as those used in Fig. 5E of Ref.10 and includes 1631 SNPs associated with acute rheumatic fever. The LD pattern in (D) involves 1787 SNPs located in the genomic region 2,012,889–2,018,654 of S. suis BM407. The pattern in (E) includes 2,205 SNPs associated with population differentiation of S. suis. The LD patterns were generated by Haploview based on the pair-wise measure of the linkage disequilibrium D’ and log likelihood of odds ratio LOD. The different LD levels are indicated in color with red for the strongest LD (D’ = 1 and LOD > 2), pink for the intermediate LD (D’ < 1 and LOD > 2) in pink, white for the weak LD (D’ < 1 and LOD < 2) in white, and purple for uninformative (D’ = 1 and LOD < 2). The average inter-SNP LD (measured as correlation coefficient r2) was significantly increased for SNPs subject to pre-selection. The between-group difference was evaluated using Wilcoxon rank-sum test

Back to article page